SQL Server Error Handling – Find if Stored Procedure Error

The error in the EXEC part of the INSERT-EXEC statement is leaving your transaction in a doomed state.

If you PRINT out XACT_STATE() in the CATCH block it is set to -1.

Not all errors will set the state to this. The following check constraint error goes through to the catch block and the INSERT succeeds.

ALTER PROCEDURE test -- or create
AS
  BEGIN try
      DECLARE @retval INT;

      DECLARE @t TABLE(x INT CHECK (x = 0))

      INSERT INTO @t
      VALUES      (1)

      SET @retval = 0;

      SELECT @retval;

      RETURN( @retval );
  END try

  BEGIN catch
      PRINT XACT_STATE()

      PRINT ERROR_MESSAGE();

      SET @retval = -1;

      SELECT @retval;

      RETURN( @retval );
  END catch; 

Adding this to the CATCH block

 IF (XACT_STATE()) = -1
BEGIN
    ROLLBACK TRANSACTION;
END;

Doesn't help. It gives the error

Cannot use the ROLLBACK statement within an INSERT-EXEC statement.

I don't think there is any way of recovering from such an error once it has happened. For your specific use case you don't need INSERT ... EXEC anyway though. You can assign the return value to a scalar variable then insert that in a separate statement.

DECLARE @RC INT;

EXEC sp_executesql
  N'EXEC @RC = test',
  N'@RC INT OUTPUT',
  @RC = @RC OUTPUT;

INSERT INTO @t
VALUES      (@RC) 

Or of course you could restructure the called stored procedure so that it doesn't raise that error at all.

DECLARE @RetVal INT = -1

IF OBJECT_ID('PrintMax', 'P') IS NULL
  BEGIN
      EXEC('create procedure PrintMax as begin print ''hello world'' end;')

      SET @RetVal = 0
  END

SELECT @RetVal;

RETURN( @RetVal ); 

The TRY...CATCH block is causing the execution of the stored procedure to terminate before it has a chance to return the values in #TEMP.

Try this stored proc:

IF OBJECT_ID('dbo.SomeProc') IS NOT NULL
DROP PROCEDURE dbo.SomeProc;
GO
CREATE PROCEDURE [dbo].[SomeProc]
AS 
BEGIN
    CREATE TABLE #TEMP
    (
        ID INT NOT NULL
    );

    INSERT INTO #TEMP 
    VALUES (1);

    INSERT INTO #TEMP 
    VALUES (1/0);

    PRINT (N'Run the SELECT');

    SELECT ID 
    FROM #TEMP;
END
GO

Running it, like this, without a TRY...CATCH, allows all the statements in the proc to run:

EXEC dbo.SomeProc;

The output:

Msg 8134, Level 16, State 1, Procedure SomeProc, Line 12
Divide by zero error encountered.
The statement has been terminated.
Run the SELECT

However, if you run it inside a TRY...CATCH:

BEGIN TRY
    EXEC dbo.SomeProc;
END TRY
BEGIN CATCH
    PRINT (ERROR_MESSAGE());
END CATCH

You see only the error message:

Divide by zero error encountered.

The PRINT (N'Run the SELECT'); never runs, and indeed the SELECT ID FROM #TEMP; never runs either. Hence no rows are returned, and nothing can be inserted into your #TEMPTABLE

From the MSDN documentation on TRY...CATCH:

If there is an error in the code that is enclosed in a TRY block, control passes to the first statement in the associated CATCH block.

As always, the devil is in the details. TRY...CATCH always aborts the code inside the BEGIN TRY...END TRY code block if any error over severity 10 occurs that does not close the database connection. Execution is immediately passed into the BEGIN CATCH...END CATCH block, even if this means aborting code in a stored proc.

Be aware that if an error occurs on any row the entire insert will not happen. The only reason you're seeing this error, and a row being inserted, is because you have two insert statements, one that runs to completion, and one that throws an error. Take for instance:

IF OBJECT_ID('dbo.SomeProc') IS NOT NULL
DROP PROCEDURE dbo.SomeProc;
GO
CREATE PROCEDURE [dbo].[SomeProc]
AS 
BEGIN
    CREATE TABLE #TEMP
    (
        ID INT NOT NULL
    );

    INSERT INTO #TEMP 
    VALUES (1);

    INSERT INTO #TEMP 
    VALUES (1)
        , (2)
        , (3)
        , (4)
        , (1/0);

    PRINT (N'Run the SELECT');

    SELECT ID 
    FROM #TEMP;
END
GO

Using this stored proc, either inside a TRY...CATCH or with no TRY...CATCH block, will only result in a single row being inserted into the #TEMPTABLE - none of the rows in the 2nd insert will be present in the output.